$1 trillion a year just to keep AI running. $40 billion in profit to cover the bill.

That staggering gap is exactly what is driving the current debate over whether we are in an AI supercycle or an AI bubble. Macro research firms like Hedgeye, alongside analysts at Goldman Sachs and Sequoia Capital, have been aggressively highlighting this exact math: the underlying unit economics of the current AI boom are deeply lopsided.

Here is a breakdown of why this massive disconnect exists and what it means for the market:

The $1 Trillion “Capex Cascade” The estimated $1 trillion a year isn’t going to end-user software; it is almost entirely infrastructure capital expenditure. Tech giants. Mainly Microsoft, Amazon, Alphabet, and Meta are in an arms race. The costs are compounding across several layers:

• The Chips: Buying tens of thousands of Nvidia GPUs.
• The Data Centers: Building the physical architecture to house them.
• The Power Grid: AI requires exponentially more electricity and specialized cooling systems than traditional cloud computing. Power infrastructure is quietly eating up a massive percentage of these budgets.

The $40 Billion Reality Check While the infrastructure spend is soaring toward $1 trillion, the actual profits generated from generative AI end-products (like ChatGPT subscriptions, enterprise AI copilots, and AI customer service bots) remain a fraction of the cost.

• The “Picks and Shovels” are Winning: Right now, the companies making the $40 billion in profit are mostly the ones selling the infrastructure (Nvidia, TSMC, Broadcom), not the companies trying to sell AI to consumers.
• The Revenue Gap: To justify a $1 trillion annual investment assuming standard tech operating margins, the AI ecosystem would eventually need to generate roughly $3 trillion in annual top-line revenue. For context, the entire global software industry currently generates roughly $1 trillion a year.

The Looming Risk: Depreciation. When tech giants plow billions into physical infrastructure, that equipment eventually ages and becomes obsolete. This creates a massive depreciation expense that sits on the balance sheet. If these companies cannot figure out how to monetize their AI models fast enough, that depreciation will begin to weigh heavily on their reported earnings in the coming years.

Firms like Hedgeye are pointing out that while AI technology is undoubtedly real, the returns on capital might not be. We are currently in the “building the railroads” phase of the technology. The risk for investors is that while the railroads changed the world, many of the original railroad investors still went bankrupt before the profitable businesses were built on top of them.

M2 Money Velocity

Is the economy booming or struggling?

Historical Baseline: From 1960 through the early 1990s, M2 velocity was relatively stable, fluctuating between 1.6 and 1.8. ( See chart above).The historical median for this metric over its full recorded history is 1.752 it is now just over 1.41.

The 2020 Collapse: The sharpest drop on the chart occurred in early 2020, where velocity plummeted to a record low of 1.103. This was driven by a massive expansion of the M2 money supply via pandemic stimulus, coupled with a sudden contraction in GDP and a freeze in everyday spending.

Current State: As of mid-2026, the ratio has modestly recovered to roughly 1.41. While this represents a steady upward trend since the pandemic lows, the current velocity remains heavily suppressed compared to historical norms.

Comparison with a Real Booming Economy

The Hallmark of a Boom: In a genuinely booming economic environment, consumer and business confidence is exceptionally high. Because individuals feel secure about their financial futures, they are much more likely to spend their cash rapidly rather than save it. This rapid exchange of currency drives the velocity of money higher, indicating frequent transactions and a highly active commercial sector.

The 1990s Benchmark: The chart illustrates a true structural boom during the late 1990s internet expansion. During this era, money was changing hands so rapidly that it drove M2 velocity to an all-time historical peak of 2.198.

The Current Stagnation: Despite standard economic narratives of recent GDP growth, the current low velocity suggests that a vast amount of money in the system is stagnating rather than circulating. Because the M2 classification encompasses less liquid assets like savings accounts, certificates of deposit, and money market funds, the suppressed velocity indicates that people and businesses are holding onto these assets for saving or investment purposes rather than deploying them for immediate consumption. In a traditionally booming economy, you would expect to see this hoarded capital unleashed into everyday transactions, driving the ratio much closer to its historical median of 1.75.

The fact that the ratio is nowhere near 1.75 indicates a slow economy. The bright outlook is the current upward trend. One wonders if the trend will continue and be sufficiently effective for the mid-term elections.

PLATINUM

OUT: On the Daily chart (29 August 2026).

URANIUM

OUT (29 August 2026) URA.

BITCOIN AND GOLD

Bitcoin outperforms Gold  (20 September 2026).  19.68 One ounce gold coins buy one Bitcoin; this was over 41 ounces in August 2025.

ETHUSD: BUY: ( 20 September 2026)

BTCUSD: BUY: 20 September 2026.  NASDAQ (Cyan).

GOLD AND SILVER STOCK

02 September 2026 Gold outperforms gold stocks.

Silver outperforms silver stocks (09 September 2026).

GOLD AND SILVER

Silver to Gold Ratio

Silver outperforming Gold. (22 July 2026)  Silver is coming out of the ground at about 8:1 and the Gold: Silver ratio is 66.52:1.

GOLD TO USD

USD outperforms Gold, (29 August 2026).

Copper (SELL: 30 August  2026)

GOLD AND SILVER CHARTS

Gold’s low of the week was $4,244.00, and the high was $4,383.00, now trading at around $4,270.00.

• Gold Monthly:  is an OUT. (25 March 2026). Looking more positive. Signals turning positive

• Gold Weekly:  (HOLD: 13 August 2026)  A very early buy signal. No reversal yet.

• Gold Daily: (OUT 02 September 2026).

The low for Silver this week was $63.30 and the high $67.64, trading around $63.63 at present. 

• Silver Monthly: (OUT: 20260430.  ( Buy $24.00 March 2024 to sell May 2026: $73.13 = 204% Gain in 2 years 2 months). A couple of indicators pointing up. No buy yet.

• Silver Weekly: (HOLD: 20 August 2026). A little weaker. No sell signals.

• Silver Daily: OUT (01 September 2026).

Clients who request an introduction to an independent precious metals supplier may contact our preferred partners by clicking here.

Any purchase or sale is conducted solely between the client and the selected supplier.

A video by Mike Adams on the Goldbacks. Well worth a look. watch here

On the Stockcharts.com charts above, the blue vertical lines are our proprietary system buy signals and the red vertical lines are system sell signals – for information purposes only

An interview with Josh, owner of Scottsdale Mint: watch here

Please contact us to arrange the purchase and storage of your gold and silver requirements in a safe, insured location outside of your jurisdiction.

If you are interested in an overview of Fort Kobbe, you may want to have a look at this video: Mike Brown, Director, Fort Kobbe, International Vaults, A DotCom Magazine Exclusive Interview

This is my interpretation of the market and is not to be taken as financial advice. Before making any buy or sell decisions, I recommend that you consult with your professional financial advisor.